Eliminate Downtime with the New 3CX SBC High Availability Clustering - Technical Preview

[saquib]

HI Seniors..

I have successfully deployed SBC HA cluster with AWS ...please guide me for monitoring ?

 

[mariosM_3CX]

HI Seniors..

I have successfully deployed SBC HA cluster with AWS ...please guide me for monitoring ?

If you mean to monitor the cluster via the Management Console, i'm afraid that this is not yet implemented.
As mentioned before, for now you can enable the Email Notification of "When the status of a trunk / SBC changes" from Management Console > Settings > Email > Notifications. With this, you'll get an email each time one of the nodes assumes the active node.

More info can be found on the published guide https://www.3cx.com/docs/sbc-high-availability-cluster/


Now if you mean to monitor the cluster directly from the SBC node machines, then there are plenty crmsh commands for this. For example you can use crm_mon to monitor the cluster status live.
More info on crmsh can be found here https://crmsh.github.io/

Later i will post here a small description on how to use crm_mon for anyone interested

 

[Splunkjamma]

Hi all,

I managed to get the HA cluster working perfectly on our 3CX system using the PI 3!. Does anyone have a guide on how to get this to work on a PI 4? For.... testing purposes

 

[Splunkjamma]

Never mind sudo apt-get update --allow-releaseinfo-change fixed the install

 

[Splunkjamma]

Unfortunately due to AWS Lightsail not allowing ICMP through their firewall HA will not work using Lightsail. You can get it to work if you open All TCP+UDP ports on the Lightsail firewall but who wants to do that?

 

[saquib]

but i have deployed lightsail ...

 

[Splunkjamma]

but i have deployed lightsail ...

How? Do you have any firewall rules set on Lightsail?

 

[saquib]

dear simple i worked ..i have creating firewall on 3cx linux system ..and lightsail firewall all ports allowed

 

[Splunkjamma]

dear simple i worked ..i have creating firewall on 3cx linux system ..and lightsail firewall all ports allowed

Ahh so you are using iptables on the actual linux system then?

 

[saquib]

Ahh so you are using iptables on the actual linux system then?

YES

 

[Splunkjamma]

YES

Nice!

 

[saquib]

Nice!

Thank you ..
dear i am interesting 3cx certification .how procedure ?

 

[Splunkjamma]

dear i am interesting 3cx certification .how procedure ?

Here you go my dude:
https://www.3cx.com/3cxacademy/

 

[saquib]

Here you go my dude:
c
[/QUOTE

 

[mariosM_3CX]

@saquib
As promised, here is a small description of how to use crm_mon to monitor the cluster status in real time. To get the following views, ssh to any of the nodes machine and type crm_mon as root user or using sudo.

While both nodes are running:


When node1 is down:


When both nodes are online, but node1 lost connectivity to the PBX:

 

[bcarrico]

I just deployed this in our office. It's working, but I did notice that "reboot phone" from the 3CX Phone menu no longer reboots the phone. Could be something here or a bug. Thought I would let you know either way. Other than that, it's working fine.

 

[mariosM_3CX]

@bcarrico If you tried the reboot shortly after the failover, this is expected.

Quoting from the guide

After a failover occurs, the newly activated SBC node can handle new calls after a few seconds. However full functionality like phone provisioning via the Management Console, can take up to two (2) minutes to be restored.

If this is the case, give a bit more time after a failover for the phones to refresh their registration and try again to reboot. If you still have this issue, we can investigate.

 

[edfosho]

Set up with two Raspberry Pi SBCs and the 3CX server in AWS, works well. I noticed (as suggested in the guide) that a network connectivity issue to one of the SBCs caused it to be configured with the virtual IP and caused confusion throughout - but after a power cycle it was fine (probably due to a service restart). So, providing power and network to the Raspberry Pi SBCs via a PoE splitter means that if it loses network (physically) it'll also get a power cycle - nice. Though, better HA cluster recovery after a logical network connectivity issue will improve this solution.

 

[cobaltit]

@mariosM_3CX Is there any additional troubleshooting that can be done, or logs gathered to help test this? It looks like we either lost internet or at least lost connectivity between our office and our cloud instance last night and our phones in the office were down this morning. crm_mon shows both sbc instances as up and running and they both agree on the same device. Wanted to know what else I can gather to help identify what is preventing this from resolving itself.

 

[JakeSt]

Out of curiosity, is there a reason 3cx went with Corosync/Pacemaker, and not something like CARP/PF/PFSync?

There's nothing quite like seeing a carp failover with synced states.